The improved abilities of collecting high volume and high quality road load data raised the complexity of analyzing them using the traditional predictive reliability techniques. Modern tools for data science are therefore required to shorten the analysis times, making it more efficient and robust.
Valeo has recently collaborated with HBM Prenscia in order to implement a new software solution for data science: nCodeDS. Starting from the so-called “Big Road Load Data”, the goal is to:
– Perform quick and efficient data post-processing
– Apply deep learning techniques to detect trends inside the data
– Build robust predictive models for reliability predictions
A case study will investigate how this approach has been used to predict in-service fatigue damage for a new range of heat exchangers.